* Regression Discontinuity sample; * This program uses data from the 1998-2001 NHIS; * that replicates the results in Card, Dobkin and Maestos; * NBER working paper number 10365. The authors use data; * for people aged 55-74 and examine the jump in insurance; * coverage at age 65 due to Medicare coverage; * this line defines the semicolon as the line delimiter; # delimit ; * there are lots of variables constructed for this program so use lots of memory; * set memork for 40 meg; set memory 40m; * write results to a log file; log using card_et_al.log,replace; *read in raw data; use card_et_al; * get list of variables; describe; * get tables of self reported health status; tab phstat; * generate some new variables; gen good_health=phstat<4; label var good_health "=1 if report,good,vgood,excel health"; * eligible for Medicare after quarter 259; gen age65=age_qtr>259; * scale the age in quarters index so that it equals 0; * in the month you become eligible for Medicare; gen index=age_qtr-260; gen index2=index*index; gen index3=index*index*index; gen index4=index2*index2; gen index_age65=index*age65; gen index2_age65=index2*age65; gen index3_age65=index3*age65; gen index4_age65=index4*age65; xi i.educ_r1 i.year i.age_qtr; * 1st stage results. Impact of Medicare on insurance coverage; * basic results in the paper. quadratic in age interacted with; * age65; reg insured male white black hispanic _Ie* _Iyear* index index2 index_age65 index2_age65 age65, cluster(index); * estimate same model but do not cluster on index. notice the drop in the std errors on ag65; reg insured male white black hispanic _Ie* _Iyear* index index2 index_age65 index2_age65 age65; * run unrestricted model that has dummies for all quarters instead of polynomial; * use for test in card/lee, equation (3); reg insured male white black hispanic _Ie* _Iyear* _Iage*; * Reduced form results. Impact of Medicare coverage on health outcomes; * Outcomes are good_health, delayed_med, not_get_med, hosp_12m; * Note that age65 is small and stat insign in all except the hospitalization equation; * implying that although Medicare has increased insurance coverage it has not changed outcomes much; reg good_health male white black hispanic _Ie* _Iyear* index index2 index_age65 index2_age65 age65, cluster(index); reg delayed_med male white black hispanic _Ie* _Iyear* index index2 index_age65 index2_age65 age65, cluster(index); reg not_get_med male white black hispanic _Ie* _Iyear* index index2 index_age65 index2_age65 age65, cluster(index); reg hosp_12m male white black hispanic _Ie* _Iyear* index index2 index_age65 index2_age65 age65, cluster(index); * examine the sensitivity of the results to the order of the polynomial in the index; * linear index; reg insured male white black hispanic _Ie* _Iyear* index age65, cluster(index); reg hosp_12m male white black hispanic _Ie* _Iyear* index age65, cluster(index); * linear index interacted with age65; reg insured male white black hispanic _Ie* _Iyear* index index_age65 age65, cluster(index); reg hosp_12m male white black hispanic _Ie* _Iyear* index index_age65 age65, cluster(index); * second order index; reg insured male white black hispanic _Ie* _Iyear* index index2 age65, cluster(index); reg hosp_12m male white black hispanic _Ie* _Iyear* index index2 age65, cluster(index); * third order index; reg insured male white black hispanic _Ie* _Iyear* index index2 index3 age65, cluster(index); reg hosp_12m male white black hispanic _Ie* _Iyear* index index2 index3 age65, cluster(index); * third order index interacted with age; reg insured male white black hispanic _Ie* _Iyear* index index2 index3 index_age65 index2_age65 index3_age65 age65, cluster(index); reg hosp_12m male white black hispanic _Ie* _Iyear* index index2 index3 index_age65 index2_age65 index3_age65 age65, cluster(index); * fourth order index; reg insured male white black hispanic _Ie* _Iyear* index index2 index3 index4 age65, cluster(index); reg hosp_12m male white black hispanic _Ie* _Iyear* index index2 index3 index4 age65, cluster(index); * fourth order index interacted with age; reg insured male white black hispanic _Ie* _Iyear* index index2 index3 index4 index_age65 index2_age65 index3_age65 index4_age65 age65, cluster(index); reg hosp_12m male white black hispanic _Ie* _Iyear* index index2 index3 index4 index_age65 index2_age65 index3_age65 index4_age65 age65, cluster(index); log close;